A selection from over 15 projects. No client names, but every figure comes from the project.
The cases come from client projects. Where that is not the case, where a system is still being built or figures are still missing, the text says so.
Timesheets from the field, phone chains, manual dispatching. An offline-capable app connects the site and the office, and it works even where typical software gives up.
A lot of knowledge was locked in people’s heads, standards and old projects, and much was done by hand. An AI platform on the company’s own server in Germany makes this knowledge usable across the whole company and checks documents before mistakes get expensive. No company knowledge leaves the building, and a person approves every suggestion. The platform is being expanded in stages; the error rate has already dropped significantly.
Inquiries from several portals landed unstructured in the inbox. The cockpit reads them in, qualifies them by clear criteria and coordinates viewing appointments, while the team keeps the approval.
Six hours a week of searching and reading across RSS feeds, news portals and case law. Some of the sources could not be reached at all. An agent scanned, filtered and rated with company context; a person gave a score that made the system better with every run.
After a listing went live, applications were screened by hand for days, followed by email back-and-forth about tenant self-disclosures, application folders and appointments. The application pre-qualifies in the background and turns the rest into one-click steps.
Applications arrived as an unstructured mass via email, portals and PDF attachments. The system extracts them, creates them in the system and pre-qualifies them, including checks for completeness and outgoing messages.
Metrics from several ad and social platforms were regularly assembled by hand. Now the exports are uploaded and the analysis is created automatically: the data plus a finished deck.
Applicant documents had to be compiled and tracked one by one. The workflows create the documents, follow up automatically and feed the project management tool.
Expert content only got written when there was time left over. The pipeline takes in notes and project results, shapes drafts in the company’s own voice and submits them for approval.
Expensive tools went missing, and nobody knew where they were. Now they get scanned instead of searched for: check-in and check-out by QR code, with a reminder when a return is overdue.
Generic outbound emails got hardly any replies. Research and outreach were automated but personalized, with multi-step follow-ups.
Applications came in by email and were copied by hand into a large spreadsheet, and funding forms were filled out from scratch every time. The system reads the email and documents, rates them in a traceable way and files everything; the forms arrive in the folder already filled out.
New contacts were researched by hand, entered into the CRM and written to one by one. Two chained workflows searched for and filtered suitable contacts, created them in the CRM and drafted an individual cover letter for each.
Across 34 processes it was unclear where AI would pay off. A scoring grid showed the best candidate. That ordering process was simplified and connected to the existing ERP system as an AI-assisted workflow, without touching its core. The case comes from the research work of one of our founders in cooperation with the group.
Receipts arrived incomplete or late, and much was entered by hand. In the application, clients upload their documents directly, an AI extracts them, and both sides work on the same data. The application is currently in a pilot.
The client had built three AI tools as prototypes themselves. They ran separately, with their own data and their own logins, and maintenance and security were not covered. We are merging them into one platform, with one login, one data foundation and a connection to the CRM. The client’s groundwork is kept. Each tool goes live on its own; figures will follow after go-live.
That is exactly what the audit is for. 45 minutes, free and no obligation. We tell you honestly what is worth it for you.
